METHOD AND SYSTEM OF DATA MODELLING

• US 20110257949A1
• Filed: 09/18/2009
• Published: 10/20/2011
• Est. Priority Date: 09/19/2008
• Status: Active Grant

• Current Assignee: University of Sydney
• Original Assignee: University of Sydney
• Inventors: Ramos, FabIo Tozeto, Vasudevan, Shrihari, Durrant-Whyte, Hugh, Nettleton, Eric
• Granted Patent: US 8,768,659 B2
• Time in Patent Office: Days
• US Class Current: 703/2
• CPC Class Codes: G06F 17/18 for evaluating statistical data , e.g. average values, frequency distributions, probability functions, regression analysis forecasting specially adapted for a specific administrative, business or logistic context G06Q10/04
• CPC Class Codes: G06N 20/10 using kernel methods, e.g. support vector machines [SVM]

Abstract

A method for modelling a dataset includes a training phase, wherein the dataset is applied to a non-stationary Gaussian process kernel in order to optimize the values of a set of hyperparameters associated with the Gaussian process kernel, and an evaluation phase in which the dataset and Gaussian process kernel with optimized hyperparameters are used to generate model data. The evaluation phase includes a nearest neighbour selection step. The method may include generating a model at a selected resolution.

Claims

1. A method for modelling data based on a dataset, comprising:

a training phase, wherein the dataset is applied to a non-stationary Gaussian process kernel in order to optimize the values of a set of hyperparameters associated with the Gaussian process kernel, and

an evaluation phase in which the dataset and Gaussian process kernel with optimized hyperparameters are used to generate model data.

2. A method as claimed in claim 1, wherein the evaluation phase comprises a nearest neighbour selection operation, wherein only a selected subset of nearest neighbour data in the dataset is used to generate each corresponding model datum.

3. A method as claimed in claim 1, wherein the Gaussian process kernel is a non-stationary neural network kernel.

4. A method as claimed in claim 1, wherein the training phase comprises a combination optimisation strategy.

5. The method of claim 4, wherein the training phase comprises simulated annealing, with the output of the simulated annealing stage being input into a Quasi-Newton optimization stage.

6. A method as claimed in claim 1 wherein the dataset has at least 100,000 members, and

wherein the method comprises sampling the dataset to obtain a training data subset that is applied in the training phase for optimization of the hyperparameters wherein the size of the training data subset is at least one order of magnitude smaller in size than the dataset.

15. The method of claim 1 when the dataset is measured terrain data.

16. The method of claim 1, comprising selecting an output resolution for the output model and generating the output model according to the selected resolution.

7-14. -14. (canceled)

17-19. -19. (canceled)

20. A method for modelling a dataset with a spatial characteristic, comprising

providing a dataset to a non-parametric, probabilistic process to derive a model from the dataset, and

saving a data set describing the model in a database that preserves the spatial characteristics of the data.

21. A method in accordance with claim 20, wherein the database utilises a KD-Tree structure to store the modelled data.

22. A method in accordance with claim 21, wherein the process utilised is a Gaussian process.

23. A method in accordance with claim 22, wherein the Gaussian process is a non-stationary Gaussian process using a neural network.

25. A method in accordance with claim 20, wherein the data set comprises hyperparameters describing the model.

26. The method of claim 20, further comprising the operation of dividing the dataset into at least two subsets, and providing each subset to the non-parametric, probabilistic process to derive a model for each of the subsets of the dataset.

24. (canceled)

27-32. -32. (canceled)

33. A system for modelling an area of terrain, comprising

a modelling module arranged to receive measured terrain data and utilise a non-parametric, probabilistic process to derive a model for the area of terrain, and

a data structure arranged to receive a data set describing the model wherein the data structure preserves the spatial characteristics of the data.

34. A system in accordance with claim 33 wherein the database utilises a KD-Tree structure to store the modelled terrain data.

35. A system in accordance with claim 34, wherein the process utilised is a Gaussian process.

36. A system in accordance with claim 35, wherein the Gaussian process is a non-stationary Gaussian process using a neural network.

37-44. -44. (canceled)

45. A computer program comprising at least one instruction arranged to implement a method for modelling data based on a dataset, comprising:

a training phase, wherein the dataset is applied to a non-stationary Gaussian process kernel in order to optimize the values of a set of hyperparameters associated with the Gaussian process kernel, and

an evaluation phase in which the dataset and Gaussian process kernel with optimized hyperparameters are used to generate model data.

46. A computer readable media incorporating a computer program comprising at least one instruction arranged to implement a method for modeling data based on a dataset, comprising:

a training phase, wherein the dataset is applied to a non-stationary Gaussian process kernel in order to optimize the values of a set of hyperparameters associated with the Gaussian process kernel, and

an evaluation phase in which the dataset and Gaussian process kernel with optimized hyperparameters are used to generate model data.